极化SAR遥感中森林特征探测

Forest characteristic detection with Pol-SAR

为了利用微波遥感数据对森林植被地区进行分类检测,寻找一种新的敏感于树木高度、种群数量和植被疏密程度等森林特征的遥感参量,该文利用中国天山地区SIR-C/X-SARL波段全极化干涉数据,对极化散射矩阵总功率、极化熵、相似性参数和极化干涉最优相关系数等极化和干涉合成孔径雷达(SAR)遥感参数进行了分析,并将这些参数进行加权组合,利用特征值分析方法求取最优加权系数,从而提出两种对森林特征敏感的参数组合表达式。在比较上述组合表达式图与同地区极化遥感总功率图之后,发现该组合表达式图中,森林地区较其他地区获得明显加强。这表明提出的组合表达式对森林地区的检测和监测具有良好的效果。

Remote sensing is used to d et ect forested area and identify forest characteristics such as tree height, bio- mass and density. SIR-C/X-SAR L-Band SLC data pairs from the Tianshan Area we re used to analyze the polarimetric and interferometric parameters, including th e total span of the scattering matrix, the polarimetric entropy, the similarity parameters and the optimal interferometric coherence in the polarimetry. The par ameters were correlated with polynomials and the optimal coefficient of the poly nomial was found using the eigenvalue matrix analysis method. The optimal polyno mial was shown to be sensitive to the forest characteristics. Comparison of the image from the optimal polynomial with that of the total span of the scattering matrix shows that the proposed method more effectively detects forested areas.